CN1164077A - Method and related apparatus for fingerprint indexing and searching - Google Patents

Method and related apparatus for fingerprint indexing and searching Download PDF

Info

Publication number
CN1164077A
CN1164077A CN 97102305 CN97102305A CN1164077A CN 1164077 A CN1164077 A CN 1164077A CN 97102305 CN97102305 CN 97102305 CN 97102305 A CN97102305 A CN 97102305A CN 1164077 A CN1164077 A CN 1164077A
Authority
CN
China
Prior art keywords
fingerprint
index
sample
subclass
indexes
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN 97102305
Other languages
Chinese (zh)
Inventor
戴勒·R·塞特拉克
约翰·C·施密特
马克·E·波斯
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Harris Corp
Harrier Inc
Original Assignee
Harrier Inc
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Harrier Inc filed Critical Harrier Inc
Priority to CN 97102305 priority Critical patent/CN1164077A/en
Publication of CN1164077A publication Critical patent/CN1164077A/en
Pending legal-status Critical Current

Links

Images

Abstract

A method for indexing and searching a plurality of reference fingerprints to determine if a sample fingerprint matches one of the reference fingerprints includes the steps of: determining respective index values for the reference fingerprints and the sample fingerprint based upon characteristics of the fingerprints and with the index values being generally evenly distributed and continuous over a relatively large population of individuals; selecting a first subset of the reference fingerprints having index values in a first range relative to the index value of the sample fingerprint; and comparing the sample fingerprint to each fingerprint of the first subset of reference fingerprints to thereby determine if there is a match between the sample fingerprint and one of the reference fingerprints and without comparing the sample fingerprint to all of the reference fingerprints.

Description

The fingerprint index is determined and search method and relevant device
The present invention relates to efficiently searching a coupling fingerprint in the bigger fingerprint database.
Fingerprint matching is the technology to a kind of reliable and widespread use of individual identification and affirmation.Specifically, the general means of fingerprint recognition relates to a sample fingerprint of scanning or its image, converts thereof into electric signal, and this finger print image and/or its specific characteristic are stored.Can be with the feature of sample fingerprint and comparing of having stored with reference to finger print information, to determine its recognition result or to reach the purpose of affirmation.
Regrettably, sample fingerprint and a large amount of comparing with reference to fingerprint may be extremely expensive and/or will spend the oversize only not time.Therefore, usually fingerprint is become a plurality of discrete set and/or subclass with the formal classification of layering tree, thereby quicken to search.For example, be that wide ridge pattern type according to fingerprint is divided into types such as dull screw thread, dull ring grain, oblique parallel be arc usually to the top classification of fingerprint.These types can also further be divided into subclass.Therefore, the sample fingerprint that search, in case itself be classified, just can be more effectively only with respective type that searches tree or subclass in those fingerprints compare.
Traditional taxonomic methods may cause carries out the branch stock with a plurality of fingerprint patterns or classification and puts, and single fingerprint base is used for fingerprint matching.Regrettably, these fingerprint bases also may be bigger, cause search time-consuming expensive.A fingerprint or be considered in some storehouses or in another storehouse, but the whether uncertainty in this or another storehouse of this fingerprint of not paying attention to are arranged again.In other words, between adjacent storehouse, do not have opening relationships usually, and this relation can help to search.The reduction of fingerprint quality also can cause it has been put into wrong storehouse simply, thereby can not obtain coupling.
When the original copy poor quality of fingerprint, this fingerprint may have many local distortions in the ridge pattern, and this can cause the misorientation of fingerprint.United States Patent (USP) 5,140, No. 642 instructions is that the basis of this method is at the method for the core point physical location of determining a fingerprint: find out ridge stream and specify a direction code, proofread and correct this ridge stream, and determine the core point position according to corrected direction code.United States Patent (USP) 5,040 has disclosed the pretreated means of a kind of fingerprint No. 224, by this pre-service, can correctly determine the core position of each fingerprint, for further mating according to the details pattern.
The details of fingerprint, for example branch of fingerprint ridge or bifurcated and terminating point all often are used to sample fingerprint and with reference to the coupling between the fingerprint database.United States Patent (USP) 3,859,633 and 3,893, No. 080 at carrying out fingerprint recognition according to Fingerprint Minutiae.
United States Patent (USP) 3,959, No. 884 instructions has disclosed a kind of fingerprint classification method, wherein fingerprint is converted to the pattern of a binary value, and this pattern has defined the space array of a ridge and paddy.
United States Patent (USP) 4,151, No. 512 instructionss have been described the method that a kind of ridge isoline data of using extraction are carried out fingerprint classification.Ridge in fingerprint pattern stream and detail data is identified and extract from the fingerprint pattern.
United States Patent (USP) 4,185, No. 270 instructionss have disclosed also coding and the affirmation process based on details.United States Patent (USP) 4,210, No. 899 instructionss disclose a kind of optical scanning fingerprint readers that have central treating stations, are used for the application of safe turnover aspect, such as allowing someone somewhither or visit to terminal is provided.United States Patent (USP) 4,525, No. 859 also at details coupling, and has described a system, and this system comprises: signal generation device is used for the every bit on the dot matrix is produced electric signal; First, second that the submatrix of lattice point is operated and the 3rd determined device; And miscellaneous part.United States Patent (USP) 4,947 has disclosed for No. 443 and a kind ofly to have utilized four features in six features to confirm and discern someone method.
Also have another kind of fingerprint matching means to attempt each fingerprint is specified the digital code of a uniqueness.United States Patent (USP) 4,747,147 instructionss have disclosed a kind of finger scan system and method, are used for rotating the one scan line around a central point on fingerprint.The code of an irregular type of representative and a ridge counting go on record together, thereby are provided the topology completely and the space description confession Computer Processing of a fingerprint by coordinate.
United States Patent (USP) 5,363,453 instructions are about producing the system and method for 24 bytes numeral fingerprint recognition sign indicating number.In addition, described the prior art device of being made by Fingermatrix company in the background parts of this invention, this device provides 400 byte digital codes.This method does not resemble the details that relies on fingerprint aforementioned many patents.
United States Patent (USP) 5,239, No. 590 instructions has disclosed a kind of finger print image disposal route, utilizes this method that a main finger print image is become a plurality of with a sample fingerprint image Segmentation, each piece is divided into a plurality of districts again, and each piece district has a plurality of picture dots again and with direction.
Method and apparatus of the present invention allow to search effectively one bigger with reference to fingerprint database, and allow to search effectively a bigger fingerprint database, the some of them fingerprint is may quality lower.
The present invention carries out the method that index is determined and searched to a plurality of with reference to fingerprint, its objective is to determine whether a sample fingerprint mates with one of reference fingerprint.In one embodiment, this method preferably includes following steps: determine with reference to fingerprint and sample fingerprint index separately that according to the feature of fingerprint its index evenly distributes on than large group and is continuous at one usually; Selection is with reference to fingerprint first subclass, and the index with reference to fingerprint wherein is in first interval of relative sample fingerprint index; With sample fingerprint with compare with reference to each fingerprint of fingerprint first subclass, thereby determine whether this sample fingerprint is complementary with reference to fingerprint with certain.Therefore, sample fingerprint does not need to compare with reference to fingerprint with all, is so in the starting stage at least, but this index determines that mode provides a kind of more effective approach to determine whether to exist and mated.
Determine preferably to determine index according to the ridge stream curvature of fingerprint with reference to the step of fingerprint and sample fingerprint index value.More particularly, the step of determining index preferably determines to curl index.This index of curling can be determined by following steps: the center of determining each fingerprint; Select the part that pre-determines of each fingerprint with respect to the center; Calculate the total amount of ridge flow path direction vector rate of change for the predetermined portions of each fingerprint.The step of selecting the fingerprint predetermined portions can be to select the part that is generally annular of each fingerprint.
According to a further aspect in the invention, select preferably to select according to the expection confidence value that sample fingerprint is arranged in first subclass with reference to the step of fingerprint first subclass.
Method of the present invention also has another important feature, and Here it is by the details of each fingerprint in the details of sample fingerprint and first subclass relatively being finished the comparison of each fingerprint in the sample fingerprint and first subclass.If determine that according to the coupling scoring its details is enough approaching, then be defined as coupling.If do not determine a coupling, this method preferably further comprises the steps: to select with reference to fingerprint second subclass, it has second index with respect to sample fingerprint that the more interval index of wide region is arranged, and then the fingerprint that increases newly in the sample fingerprint and second subclass is compared.This process can repeat, up to having found out a coupling or till the distribution situation of index shows when can not find coupling probably.
The present invention also obtains a result, and promptly the orientation of index and fingerprint, position and yardstick are irrelevant.In addition, reference fingerprint and sample fingerprint are determined that the step of index comprises that the finger print information that is produced by comparison step determines this index that calculates.For example, in the place that details in fingerprint is used for comparison step, the processing of extracting details also can produce according to one embodiment of present invention and the ridge that uses flows curvature.
Whether its equipment is to be used for index to determine and search a plurality ofly with reference to fingerprint mate with one of reference fingerprint to determine a sample fingerprint.This equipment comprises that index determines device, is used for determining with reference to fingerprint and sample fingerprint index separately according at least one predetermined fingerprint characteristic, and index is generally evenly distributed on than large group one.This equipment also comprises selecting arrangement, be used to select first subclass with reference to fingerprint, its index falls in first interval with respect to the sample fingerprint index, comprise that also comparison means is used for comparing with sample fingerprint with reference to each fingerprint of fingerprint first subclass, thereby determine whether to exist sample fingerprint and with reference to the coupling between one of fingerprint, and needn't require at the very start sample fingerprint and all are compared with reference to fingerprint.
Index is determined that device comprises and is determined with reference to the go forward side by side device of line ordering of the index of fingerprint and sample fingerprint, make these index monotone variation and generally change continuously on than large group one, for its index of the fingerprint that repeats of a finger of body one by one generally should be stable, index generally is increment type to be changed along with the decline of fingerprint quality, index can be by being compared that device produces or coming out for the information calculations of comparison means generation, and/or generally do not rely on fingerprint orientation, position and yardstick.
Selecting arrangement can comprise that device is used for selecting according to the expectation confidence value that sample fingerprint is positioned at first subclass first subclass of sample for reference.Comparison means can comprise that preferably device is used for the details of each fingerprint of the details of sample fingerprint and first subclass is compared.Selecting arrangement can further include a device, if with the comparison of fingerprint first subclass in do not find coupling, then available this device is selected with reference to fingerprint second subclass, this second subclass has the interval index of broad with respect to the index of sample fingerprint.
Index determines that device can comprise a device, is used for the definite index of curling with reference to fingerprint and sample fingerprint of ridge stream according to fingerprint.More particularly, index determines that device can comprise: the center determines that device is used for determining the center of each fingerprint, the choice of location device is used to select the predetermined portions of fingerprint with respect to the center, and calculation element is used for the total amount of calculated fingerprint reservations costula flow path direction vector rate of change.
Now will by way of example and with reference to the accompanying drawings the present invention be described, in the accompanying drawing:
Fig. 1 is the schematic block diagram of this equipment one embodiment;
Fig. 2 is the schematic block diagram of another embodiment part of this equipment;
Fig. 3 is the process flow diagram of this method part of explanation;
Fig. 4 is the process flow diagram of this method another part of explanation;
Fig. 5 is the schematic plan view as in one embodiment fingerprint part;
Fig. 6 is the schematic theoretical diagram of curling index in one embodiment;
Fig. 7 is the ridge flow path direction vector and the horizontal component figure thereof of a sample fingerprint;
Fig. 8 is the ridge flow path direction vector and the vertical component figure thereof of a sample fingerprint;
Fig. 9 is the ridge flow path direction vector of a sample fingerprint and total ridge flow point spirogram thereof; And
Figure 10 and Figure 11 are the index map of describing in each example of curling.
At first the present invention is described with reference to figure 1 and Fig. 2.The equipment 20 of Fig. 1 comprises a fingerprint scanner or sensor 22, is used to import the image of a fingerprint.A suitable electric field measurement type fingerprint sensor 22 is described in the unsettled U.S. Patent application that is entitled as " electric field fingerprint sensor and correlation technique " (sequence number 08/592469).Fingerprint sensors optics, ultrasonic and other types also can adopt.In addition, finger print image can be from the paper copy or the slide input of fingerprint.
Finger print image is handled by device of detail extraction shown in the figure 24 and crimpness index determiner 26, will explain in more detail hereinafter.A plurality ofly determined by index and store in the database 25 shown in the figure with reference to fingerprint.The detail data of being determined by index in the database 25 can selectively be searched by the comparer of details shown in the figure 27.The selector switch of level of confidence shown in the figure 26 provides the suitable interval index that initially searches.Details comparer 27 can provide one " coupling " or " not matching " output, as shown in FIG..
Its output can be used for the identification individual, as allowing visit or allowing the individual to finish some further function.When determining whether someone had before registered in this database, also wishing has this coupling or the output that do not match.
Be appreciated that another modification according to equipment of the present invention with reference to figure 2, wherein equipment 30 comprises a plurality of parallel processor 31a-31n that arrange by cooperative relationship.Processor 31a-n can designated specific searching between the index area, thereby the total efficiency of equipment 30 and operating speed are increased, and surpasses when using separately the means that search based on traditional classification.
Specifically can further understand overall operation of the present invention with reference to the process flow diagram of figure 3 and Fig. 4.From section start (piece 50), in the database each with reference to the fingerprint extraction details in fingerprint and calculate the crimpness index, and at piece 52 place's event memories.The another kind of practice perhaps as the replenishing of the above-mentioned practice, also can be stored and searches after each digitized picture with reference to fingerprint is provided with; Yet, wish that image and corresponding crimpness index or other are searched index to be stored together, as this paper will be in greater detail.
As shown in the piece 54, an input or sample fingerprint have been obtained.Extract details and calculate the crimpness index in 56 pairs of sample fingerprint of piece.Can be index first interval selected level of confidence of wanting to search at piece 58.For example, the user can and search the balance of required time between the two based on the possibility that finds coupling and select level of confidence.Therefore, can to represent these indexes be center or the distribution that is associated with the sample fingerprint index with the sample fingerprint index in index first interval.Have, for a bigger database, this interval can be the index of sample fingerprint again.
At piece 60, compare with reference to first subclass of fingerprint and input or sample fingerprint, to determine whether to exist a coupling at piece 62.For example, if details coupling scoring is in advance in the interval of being selected by the user, just can be defined as coupling.This details coupling is widely used in other traditional fingerprints and searches in the application, so be easy to need not further discussion here into the one's own profession expert understands.
If be defined as a coupling, then can point out a coupling, and another is searched this process that repeats at piece 66 at piece 62.If do not find coupling, then can select second subclass that contains wideer interval index for further searching shown in the piece 64 at piece 62.Because according to the hope feature of index of the present invention, contiguous index more likely produces a coupling than index at a distance.
Now more specifically turn to the process flow diagram of Fig. 4, and Fig. 5 and Fig. 6, to further describe crimpness index according to an embodiment of the invention.At initial (piece 70) afterwards, can select the part of fingerprint for further analysis at piece 72.As shown in Figure 5, can select annulus according to the central point 83 of fingerprint ridge 84 by radius r 1 and r2 definition.Can easily determine the central point 83 of fingerprint by traditional algorithm and means.For example, can use United States Patent (USP) 5,140, No. 642 and 5,040, No. 224 described means of United States Patent (USP) are come centre of location point.
Can calculate ridge flow path direction vector at piece 74.Can be randomly in the size of piece 76 compute vectors, and calculate the total amount of ridge flow path direction vectors, and before piece 80 places stop, determining the crimpness index at piece 78.Higher crimpness index shows that the ridge flow path direction changes greatly, promptly has more significant vortex pattern.
The calculating of crimpness index can be expressed as following formula: ξ = 1 q ∫ ∫ Δθ ( x ′ , y ′ ) ∂ x ∂ y
r 1 2≤ x ' 2+ y ' 2≤ r 2 2Here:
ζ=crimpness index
θ=local ridge deflection
r 1The inside radius in=valuation zone
r 2The external radius in=valuation zone
The valuation zone number percent of the available ridge directional data of q=fingerprint quality index=contain
X ' and y ' are with respect to the normalized coordinate of fingerprint pattern computing center.
Refer again to the synoptic diagram of Fig. 6 now, other similar indexes of crimpness exponential sum are further explained.When each fingerprint in the database is calculated the crimpness index, just can be used for fingerprint is lined up monotone increasing sequence as figure 85 representatives.The crimpness index that sample fingerprint is calculated drops on the position of being found the coupling fingerprint by most probable in the data queued storehouse.By those inquiries are directed to the zone that those most probables at first find coupling, thereby improved the performance that searches.Adaptation can be from being begun to search by the assigned address the data queued storehouse.If that fingerprint is failed coupling, for example mark and indicate by not satisfying predetermined details match reference, then adaptation is attempted by the top or following fingerprint that is close in the data queued storehouse.In other words, search second subclass that can be extended to index.This process will repeat to be found till the coupling, perhaps to the possibility of coupling extremely low so that have no reason or when not wishing to continue to search again till.
The crimpness index that obtains according to the present invention has many expected characteristicses that help efficiently and accurately to search.The present invention has considered that also other indexes determine method.Ridge stream second derivative total amount can be used as index.Another kind method is to produce point of crossing pattern by each fingerprint of taking from each finger of the individual left hand and the right hand, and these point of crossing can be used as a kind of index and determine method, and has one or more in the desired characteristic that is realized by the crimpness index.
This index preferably is easy to line up to define a function that is generally monotone variation, and is also preferably continuous, and is stable to this index of repetition fingerprint of individual finger.Wish that also increment type changes these indexes along with the decline of fingerprint quality usually.Therefore, if fingerprint quality has decline slightly, then index only changes a less amount.Like this, (this may be common situation) determines with querying method very big superiority is arranged just according to index of the present invention when all fingerprints may all have quality to descend slightly.In this case, the index range that search is associated with the expectation confidence value that finds a coupling, thereby can consider that the quality of fingerprint selects this index range, and this is very beneficial.
This index does not also rely on orientation, position and the yardstick of fingerprint.Have again, preferably can calculate according to the finger print information that produces for comparison step to reference fingerprint and sample fingerprint index.For example, use the place of details in fingerprint in comparison step, the processing procedure of extracting details in crimpness index embodiment of the present invention also can produce previously described ridge flow data.Index according to this method can be an one dimension, but the exponentiation algorithm of two dimension or multidimensional also is possible.Also can be used for the characteristic that index is determined to the step that compares with reference to fingerprint in the sample fingerprint and first subclass based on the fingerprint characteristic that does not rely on ridge stream or other.
The sample fingerprint that is searched has the situation of a coupling to have many jumbo identifications in database and authorizes and use.In these situations, the user registers in database and expectation is recognized.This class is used and can be obtained an advantage from crimpness index embodiment according to the present invention and other embodiment especially.The situation that searching fingerprint does not have to mate in database is a kind of situation of initial registration, at this moment uses the overall situation to search and identifies the people who re-registers with another name, also has in many enforcements registration application this situation is also arranged.
With reference to figure 7-9, explain that further the representative crimpness to sample fingerprint calculates.In Fig. 7, demonstrate the tabulation that a sample fingerprint ridge flow path direction vector and level thereof are curled component by matrix 86.Angled line among the figure in each ridge stream unit 89 is represented ridge direction stream.Numeral 88 representatives in each unit 89 are curled component in the level of the calculating of that position.Similarly, Fig. 8 shows and vertically curls component, Fig. 9 reveal competence and vertical total amount of curling component.Use the analogous element among apostrophe and two apostrophe difference presentation graphs 8 and Fig. 9.In these illustrated exemplary, used the fingerprint core.Certainly, as mentioned before, may not use the zone that is in the fingerprint center in other embodiments of the invention.
The several times test result of crimpness index has confirmed that use crimpness index can significantly improve the performance that fingerprint searches (cold-search) operation.
At two finger print data library tests that can utilize the crimpness index, search the ability that the desired data storehouse searches performance to determine this index to improving fingerprint.
First database is the direct scanning fingerprint that is obtained by ultrasonic fingerprint scanner.This database comprises the data of 103 fingers, and each finger has two representative images, represents with square and rhombus in Figure 10, and constitute figure 91, this is to draw in proper order according to the queuing of the crimpness exponential average of each finger.It is sliding not carry out the ridge levelling in the used crimpness algorithm, and it is fuzzy indefinite that all fingerprints all have the ridge stream unit of 5-15%.First database hereto, its performance evaluation shows:
1. when searching fingerprint arranged 1 when coupling in database, the average behavior that the overall situation is searched was improved to 15: 1.
2. not existing the overall situation of coupling to search improvement in performance with the affirmation of 99% degree of confidence is 5: 1.
Second database is the direct scanning fingerprint of 81 fingers, and each finger has 5 sampling images.These data are lined up, and by the figure among Figure 11 92 expression, wherein the index of 5 of each finger images is respectively by along the geometirc graphical presentation of perpendicular line separately.The performance evaluation of carrying out at this database shows:
1. when searching fingerprint arranged 1 when coupling in database, the average behavior that the overall situation is searched was improved to 10: 1.
2. not existing the overall situation of coupling to search improvement in performance with the affirmation of 95% degree of confidence is 3: 1.
And with reference to the method for one of fingerprint coupling whether index determine and search a plurality ofly to comprise the steps: to determine with reference to fingerprint and sample fingerprint index separately according to the feature of fingerprint to determine a sample fingerprint with reference to fingerprint, and its index evenly distributes on than large group and is continuous at one usually; Selection is with reference to fingerprint first subclass, and the index with reference to fingerprint wherein is in first interval of relative sample fingerprint index; With sample fingerprint with compare with reference to each fingerprint of fingerprint first subclass, thereby determine whether this sample fingerprint is complementary with reference to fingerprint with certain, and do not need to compare with reference to fingerprint with all.Reference fingerprint and sample fingerprint are determined the step of index preferably determines index according to the ridge stream curvature of fingerprint.

Claims (17)

  1. One kind to a plurality of with reference to fingerprint carry out index determine and search with determine a sample fingerprint whether with the method for one of reference fingerprint coupling, this method comprises the steps:
    Determine with reference to fingerprint and sample fingerprint index separately according at least one predetermined fingerprint feature, and index is generally evenly distributed on than large group at one;
    Selection is with reference to fingerprint first subclass, and these indexes with reference to fingerprint are in first interval for the sample fingerprint index; And
    With sample fingerprint and relatively, with reference to each fingerprint in fingerprint first subclass thus determine whether to exist sample fingerprint and mate with reference to one of fingerprint.
  2. 2. one kind is carried out the method that index is determined and searched to a plurality of with reference to fingerprint, comprises the steps:
    Determine with reference to fingerprint and sample fingerprint index separately according at least one predetermined fingerprint feature, and index is generally evenly distributed on than large group at one; And
    For searching, these indexes with reference to fingerprint are in first interval with respect to the sample fingerprint index with reference to fingerprint first subclass in selection.
  3. 3. method as claimed in claim 1 or 2 is wherein determined to reference fingerprint and sample fingerprint that the step of index comprises and is determined and get rid of these indexes, makes it become the function that is generally monotone variation on than large group at.
  4. 4. as the described method of one of claim 1 to 3, wherein reference fingerprint and sample fingerprint are determined that the step of index comprises and determine these indexes that making it is continuous at one on than large group.
  5. 5. as the described method of one of claim 1 to 4, wherein reference fingerprint and sample fingerprint are determined that the step of index comprises definite these index, make it be increment along with the decline of fingerprint quality and change.
  6. 6. as the described method of one of claim 1 to 5, wherein reference fingerprint and sample fingerprint are determined that the step of index comprises definite these indexes, makes it generally not rely on fingerprint orientation, position and yardstick.
  7. 7. as the described method of one of claim 1 to 6, wherein select step with reference to fingerprint first subclass to comprise to be arranged in reference to the expectation confidence value of fingerprint first subclass to select with reference to fingerprint first subclass according to sample fingerprint.
  8. 8. as the described method of one of claim 1 to 7, wherein definite step with reference to fingerprint and sample fingerprint index comprises according to the ridge stream feature of fingerprint determines these indexes.
  9. 9. as the described method of one of claim 1 to 8, wherein determine to form by following steps with reference to the step of fingerprint and sample fingerprint:
    Determine the center of each fingerprint;
    Select a predetermined portions with respect to the center; And
    The predetermined portions of each fingerprint is calculated the total amount of ridge direction vector rate of change.
  10. An index determine and search a plurality of with reference to fingerprint to determine a sample fingerprint whether with reference to the equipment of one of fingerprint coupling, this equipment comprises:
    Index is determined device, is used for determining with reference to fingerprint and sample fingerprint index separately according at least one predetermined fingerprint feature, and these indexes are generally evenly distributed on than large group at one;
    Selecting arrangement is used to select first subclass with reference to fingerprint, and the index of these fingerprints is in first interval with respect to the index of sample fingerprint; And
    Comparison means is used for comparing with sample fingerprint with reference to each fingerprint of fingerprint first subclass, with Hu Ding sample fingerprint and mate with reference to one of fingerprint whether.
  11. 11. as equipment as described in the claim 10, wherein said index determines that device comprises and determine and the device of these indexes of lining up, making these indexes is the function of monotone variation one on than large group substantially.
  12. 12. as claim 10 or 11 described equipment, wherein said index determines that device comprises the device of determining with reference to fingerprint and sample fingerprint index, makes these indexes generally continuous on than large group one.
  13. 13. as the described equipment of one of claim 10 to 12, wherein said index determines that device comprises the device of determining with reference to fingerprint and sample fingerprint index, and these indexes can be calculated from produced the information for use by comparison means.
  14. 14. equipment according to claim 10, wherein said selecting arrangement comprise the device of selection with reference to fingerprint first subclass, this selection is a basis with the expectation confidence value that sample fingerprint is arranged in reference to fingerprint first subclass; Also comprise the details of sample fingerprint and the device that compares with reference to the details of each fingerprint in fingerprint first subclass.
  15. 15. equipment as claimed in claim 14, wherein said selecting arrangement comprises that also this second subclass has than first interval roomy index second interval if do not find coupling then select device with reference to fingerprint second subclass in to the comparison of reference fingerprint first subclass.
  16. 16. as the described equipment of one of claim 10 to 15, wherein said index determines that device comprises according to the definite device with reference to fingerprint and sample fingerprint index of fingerprint ridge stream.
  17. 17. equipment as claimed in claim 16, wherein said index determines that device comprises:
    Device is determined at the center, is used for determining the center of each fingerprint;
    The part selecting arrangement is used for the part with respect to center selection fingerprint; And
    Calculation element is used for the total amount to the predetermined portions calculating ridge direction vector rate of change size of each fingerprint.
CN 97102305 1996-01-23 1997-01-20 Method and related apparatus for fingerprint indexing and searching Pending CN1164077A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN 97102305 CN1164077A (en) 1996-01-23 1997-01-20 Method and related apparatus for fingerprint indexing and searching

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US589064 1996-01-23
CN 97102305 CN1164077A (en) 1996-01-23 1997-01-20 Method and related apparatus for fingerprint indexing and searching

Publications (1)

Publication Number Publication Date
CN1164077A true CN1164077A (en) 1997-11-05

Family

ID=5166237

Family Applications (1)

Application Number Title Priority Date Filing Date
CN 97102305 Pending CN1164077A (en) 1996-01-23 1997-01-20 Method and related apparatus for fingerprint indexing and searching

Country Status (1)

Country Link
CN (1) CN1164077A (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102842042A (en) * 2011-06-07 2012-12-26 埃森哲环球服务有限公司 Biometric authentication technology
CN106716442A (en) * 2014-09-24 2017-05-24 久田康弘 Fingerprint authentication system, fingerprint authentication program, and fingerprint authentication method
CN109034815A (en) * 2017-06-09 2018-12-18 阿里巴巴集团控股有限公司 The method and device of safety verification is carried out based on biological characteristic

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102842042A (en) * 2011-06-07 2012-12-26 埃森哲环球服务有限公司 Biometric authentication technology
CN102842042B (en) * 2011-06-07 2015-04-22 埃森哲环球服务有限公司 Biometric authentication technology
CN106716442A (en) * 2014-09-24 2017-05-24 久田康弘 Fingerprint authentication system, fingerprint authentication program, and fingerprint authentication method
CN109034815A (en) * 2017-06-09 2018-12-18 阿里巴巴集团控股有限公司 The method and device of safety verification is carried out based on biological characteristic
CN109034815B (en) * 2017-06-09 2021-06-18 创新先进技术有限公司 Method and device for performing security verification based on biological characteristics
US11954190B2 (en) 2017-06-09 2024-04-09 Advanced New Technologies Co., Ltd. Method and apparatus for security verification based on biometric feature

Similar Documents

Publication Publication Date Title
US5845005A (en) Apparatus for fingerprint indexing and searching
JP3689455B2 (en) Information processing method and apparatus
US6178417B1 (en) Method and means of matching documents based on text genre
AU706080B2 (en) Character recognition system identification of scanned and real time handwritten characters
US6687401B2 (en) Pattern recognizing apparatus and method
US6621941B1 (en) System of indexing a two dimensional pattern in a document drawing
EP0567680B1 (en) Pattern recognition and validation, especially for hand-written signatures
CN1136848A (en) Fingerprint matching system
US6321232B1 (en) Method for creating a geometric hash tree in a document processing system
US20070168382A1 (en) Document analysis system for integration of paper records into a searchable electronic database
CN1163841C (en) On-line hand writing Chinese character distinguishing device
EP0355748A2 (en) A pattern recognition apparatus and method for doing the same
US5438628A (en) Method for matching text images and documents using character shape codes
WO1993020533A1 (en) Character-recognition systems and methods with means to measure endpoint features in character bit-maps
CN1106620C (en) Information processing method and apparatus
CN100390815C (en) Template optimized character recognition method and system
JP3917349B2 (en) Retrieval device and method for retrieving information using character recognition result
CN1164077A (en) Method and related apparatus for fingerprint indexing and searching
EP1010128B1 (en) Method for performing character recognition on a pixel matrix
Shtok et al. CHARTER: heatmap-based multi-type chart data extraction
JP3666909B2 (en) Character recognition apparatus and method
Marcelli et al. A structural indexing method for character recognition
CA2617416C (en) Character recognition system identification of scanned and real time handwritten characters
MARCELLI et al. A MULTI-LEVEL DESCRIPTION OF CHARACTER SHAPE®
CN117149947A (en) Book query method and device based on VR equipment, VR equipment and medium

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C02 Deemed withdrawal of patent application after publication (patent law 2001)
WD01 Invention patent application deemed withdrawn after publication